A GPU-Based Enhanced Genetic Algorithm for Power-Aware Task Scheduling Problem in HPC Cloud
In this paper, we consider power-aware task scheduling (PATS) in HPC clouds. Users request virtual machines (VMs) to execute their tasks. Each task is executed on one single VM, and requires a fixed number of cores (i.e., processors), computing power (mil
- PDF / 631,757 Bytes
- 11 Pages / 439.363 x 666.131 pts Page_size
- 88 Downloads / 236 Views
Abstract. In this paper, we consider power-aware task scheduling (PATS) in HPC clouds. Users request virtual machines (VMs) to execute their tasks. Each task is executed on one single VM, and requires a fixed number of cores (i.e., processors), computing power (million instructions per second - MIPS) of each core, a fixed start time and non-preemption in a duration. Each physical machine has maximum capacity resources on processors (cores); each core has limited computing power. The energy consumption of each placement is measured for cost calculating purposes. The power consumption of a physical machine is in a linear relationship with its CPU utilization. We want to minimize the total energy consumption of the placements of tasks. We propose here a genetic algorithm (GA) to solve the PATS problem. The GA is developed with two versions: (1) BKGPUGA, which is an adaptively implemented using NVIDIA’s Compute Unified Device Architecture (CUDA) framework; and (2) SGA, which is a serial GA version on CPU. The experimental results show the BKGPUGA program that executed on a single NVIDIA® TESLA™ M2090 GPU (512 cores) card obtains significant speedups in comparing to the SGA program executing on Intel® XeonTM E5-2630 (2.3 GHz) on same input problem size. Both versions share the same GA’s parameters (e.g. number of generations, crossover and mutation probability, etc.) and a relative small (10-11) on difference of two finesses between BKGPUGA and SGA. Moreover, the proposed BKGPUGA program can handle large-scale task scheduling problems with scalable speedup under limitations of GPU device (e.g. GPU’s device memory, number of GPU cores, etc.).
1
Introduction
Cloud platforms have become more popular in provision of computing resources under virtual machine (VM) abstraction for high performance computing (HPC) users to run their applications. An HPC cloud is such a cloud platform. Keqin Li [1] presented a task scheduling problems and power-aware scheduling algorithms on multiprocessor computers. We consider here the power-aware task scheduling (PATS) problem in the HPC cloud. The challenge of the PATS problem is the trade-off between minimizing of energy consumption and satisfying Quality of Service (QoS) (e.g. performance or on-time resource availability for reservation requests). Genetic algorithm (GA) has proposed to solve task scheduling problems [2]. Moreover, GA is one of evolutionary Linawati et al. (Eds.): ICT-EurAsia 2014, LNCS 8407, pp. 159–169, 2014. © IFIP International Federation for Information Processing 2014
160
Q.-H. Nguyen et al.
inspired algorithms that are used in green computing [3]. The PATS problem with N tasks (each task requires a VM) and M physical machines can generate MN possible placements. Therefore, whenever the PATS problem increases its problem size, the computation time of these algorithms to find out an optimal solution or a satisfactory solution is unacceptable. GPU computing has become a popular programming model to get high performance on data-parallel applications. NVIDIA introd
Data Loading...
